scoruby 0.2.1 → 0.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: e090c4502c064f07dc851a28873a6e60bbd7a301
-  data.tar.gz: 3ef721448831698663309efa187103971ebe8ce1
+  metadata.gz: 29d319117ac389447c955bcd448235865fd72104
+  data.tar.gz: 851feb27437019ef236b7ab8130c52800f3239a4
 SHA512:
-  metadata.gz: da34ae5f3501a5f73cf28c43bd4d736f07fda38988c0f4c4504c8418a4f9e990e9a97eb94afbd2af405001252100c3b96ab43c1a61f102083b19f5725aa90714
-  data.tar.gz: e038681f58084d1c80998d4d6075969917072c6e45b52c67427311e9b304e85a9f9d8d560e99d20be3e0045d89ce2afcdee0bd37509dc9b446caff998b06fa30
+  metadata.gz: e13cff01a0cf6800aa30c522dbb27bee0cfa931f55c7aa466811f7a46aff6fc865d9d9cdcacd2ed462b334cd6e38078c6e92a5235287cadd62383ae8abf1ff22
+  data.tar.gz: b8664649eeef47a290f0c8547b486896d0044b3803580639468d41f8f1098e27c4de8b5f0df06a8807463e39c8d827e101fac4457de1ef3dcee947887404fe37

data/README.md

@@ -140,6 +140,28 @@ gbm.score(features)
 
 ```
 
+### Decision Tree
+
+#### Classify by PMML - Ruby
+
+```ruby
+decision_tree = Scoruby.get_model 'decision_tree.pmml'
+
+features =  {
+        Sex: 'male',
+        Parch: 0,
+        Age: 30,
+        Fare: 9.6875,
+        Pclass: 2,
+        SibSp: 0,
+        Embarked: 'Q'       
+    }
+
+decision_tree.decide(features)
+
+=> #<Decision:0x007fc232384180 @score="0", @score_distribution={"0"=>"0.999615579933873", "1"=>"0.000384420066126561"}>
+```
+
 ## Development
 
 After checking out the repo, run `bin/setup` to install dependencies. Then, run `rake rspec` to run the tests. You can also run `bin/console` for an interactive prompt that will allow you to experiment.

data/lib/{random_forest → models}/gbm.rb

@@ -1,4 +1,4 @@
-require 'decision_tree'
+require 'models/decision_tree'
 require 'features'
 
 class Gbm

data/lib/{random_forest → models}/random_forest.rb

@@ -1,4 +1,4 @@
-require 'decision_tree'
+require 'models/decision_tree'
 
 class RandomForest
   RF_FOREST_XPATH = 'PMML/MiningModel/Segmentation/Segment'

data/lib/models_factory.rb

@@ -0,0 +1,28 @@
+require 'models/random_forest'
+require 'models/gbm'
+
+class ModelsFactory
+  RANDOM_FOREST_MODEL = 'randomForest_Model'
+  GBM_INDICATION = '//OutputField[@name="scaledGbmValue"]'
+  MODEL_NOT_SUPPORTED_ERROR = 'model not supported'
+
+  def self.factory_for(xml)
+    return RandomForest.new(xml) if random_forest?(xml)
+    return Gbm.new(xml) if gbm?(xml)
+    return DecisionTree.new(xml.child) if decision_tree?(xml)
+
+    raise MODEL_NOT_SUPPORTED_ERROR
+  end
+
+  def self.decision_tree?(xml)
+    !xml.xpath('PMML/TreeModel').empty?
+  end
+
+  def self.random_forest?(xml)
+    xml.xpath('PMML/MiningModel/@modelName').to_s == RANDOM_FOREST_MODEL
+  end
+
+  def self.gbm?(xml)
+    !xml.xpath(GBM_INDICATION).empty?
+  end
+end

data/lib/{random_forest/predicate_factory.rb → predicate_factory.rb}

@@ -1,6 +1,8 @@
-require 'compound_predicate'
-require 'simple_predicate'
-require 'simple_set_predicate'
+require 'predicates/compound_predicate'
+require 'predicates/simple_predicate'
+require 'predicates/simple_set_predicate'
+require 'predicates/true_predicate'
+require 'predicates/false_predicate'
 
 class PredicateFactory
 
@@ -13,23 +15,4 @@ class PredicateFactory
   end
 end
 
-class TruePredicate
-  def field
-    nil
-  end
-
-  def true?(_)
-    true
-  end
-end
-
-class FalsePredicate
-  def field
-    nil
-  end
-
-  def true?(_)
-    false
-  end
-end
 

data/lib/{random_forest → predicates}/compound_predicate.rb

@@ -1,5 +1,4 @@
 class CompoundPredicate
-  # TODO: spec
 
   attr_reader :field
 
@@ -11,7 +10,7 @@ class CompoundPredicate
     @predicates = []
     @predicates << PredicateFactory.for(children[0])
     @predicates << PredicateFactory.for(children[1])
-    @field = @predicates.map(&:field)
+    @field = @predicates.map(&:field).flatten.compact
   end
 
   def true?(features)
@@ -20,10 +19,14 @@ class CompoundPredicate
     and?(features) if @boolean_operator == 'and'
   end
 
+  def is_missing?(features)
+    @field.any? { |f| !features.keys.include?(f) }
+  end
+
   private
 
   def surrogate?(features)
-    # TODO: return 1 if 0 is missing
+    return @predicates[1].true?(features) if @predicates[0].is_missing?(features)
     @predicates[0].true?(features)
   end
 

data/lib/predicates/false_predicate.rb

@@ -0,0 +1,13 @@
+class FalsePredicate
+  def field
+    nil
+  end
+
+  def true?(_)
+    false
+  end
+
+  def is_missing?(_)
+    false
+  end
+end

data/lib/{random_forest → predicates}/simple_predicate.rb

@@ -25,6 +25,12 @@ class SimplePredicate
     features[field].nil? || !features.has_key?(field) if @operator == IS_MISSING
   end
 
+  def is_missing?(features)
+    !features.keys.include?(@field)
+  end
+
+  private
+
   def num_true?(features)
     return false unless features[@field]
     curr_value = Float(features[@field])

data/lib/{random_forest → predicates}/simple_set_predicate.rb

@@ -15,6 +15,10 @@ class SimpleSetPredicate
     @array.include? features[@field] if @operator == IS_IN
   end
 
+  def is_missing?(features)
+    !features.keys.include?(@field)
+  end
+
   private
 
   def single_or_quoted_words(string)

data/lib/predicates/true_predicate.rb

@@ -0,0 +1,13 @@
+class TruePredicate
+  def field
+    nil
+  end
+
+  def true?(_)
+    true
+  end
+
+  def is_missing?(_)
+    false
+  end
+end

data/lib/scoruby/version.rb

@@ -1,3 +1,3 @@
 module Scoruby
-  VERSION = '0.2.1'
+  VERSION = '0.2.2'
 end

data/lib/scoruby.rb

@@ -1,15 +1,10 @@
 require 'scoruby/version'
+require 'models_factory'
 require 'nokogiri'
-require 'random_forest'
-require 'gbm'
 require 'logger'
 require 'pry'
 
 module Scoruby
-  RANDOM_FOREST_MODEL = 'randomForest_Model'
-  GBM_INDICATION = '//OutputField[@name="scaledGbmValue"]'
-  MODEL_NOT_SUPPORTED_ERROR = 'model not supported'
-
   class << self
     attr_writer :logger
 
@@ -22,13 +17,7 @@ module Scoruby
 
   def self.get_model(pmml_file_name)
     xml = xml_from_file_path(pmml_file_name)
-    new_model(xml)
-  end
-
-  def self.new_model(xml)
-    return RandomForest.new(xml) if random_forest?(xml)
-    return Gbm.new(xml) if gbm?(xml)
-    raise MODEL_NOT_SUPPORTED_ERROR
+    ModelsFactory.factory_for(xml)
   end
 
   def self.xml_from_file_path(pmml_file_name)
@@ -40,12 +29,4 @@ module Scoruby
     xml = Nokogiri::XML(pmml_string) { |config| config.noblanks }
     xml.remove_namespaces!
   end
-
-  def self.random_forest?(xml)
-    xml.xpath('PMML/MiningModel/@modelName').to_s == RANDOM_FOREST_MODEL
-  end
-
-  def self.gbm?(xml)
-    !xml.xpath(GBM_INDICATION).empty?
-  end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: scoruby
 version: !ruby/object:Gem::Version
-  version: 0.2.1
+  version: 0.2.2
 platform: ruby
 authors:
 - Asaf Schers
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2017-07-07 00:00:00.000000000 Z
+date: 2017-07-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -124,16 +124,19 @@ files:
 - Rakefile
 - bin/console
 - bin/setup
-- lib/random_forest/compound_predicate.rb
-- lib/random_forest/decision.rb
-- lib/random_forest/decision_tree.rb
-- lib/random_forest/features.rb
-- lib/random_forest/gbm.rb
-- lib/random_forest/node.rb
-- lib/random_forest/predicate_factory.rb
-- lib/random_forest/random_forest.rb
-- lib/random_forest/simple_predicate.rb
-- lib/random_forest/simple_set_predicate.rb
+- lib/decision.rb
+- lib/features.rb
+- lib/models/decision_tree.rb
+- lib/models/gbm.rb
+- lib/models/random_forest.rb
+- lib/models_factory.rb
+- lib/node.rb
+- lib/predicate_factory.rb
+- lib/predicates/compound_predicate.rb
+- lib/predicates/false_predicate.rb
+- lib/predicates/simple_predicate.rb
+- lib/predicates/simple_set_predicate.rb
+- lib/predicates/true_predicate.rb
 - lib/scoruby.rb
 - lib/scoruby/version.rb
 - scoruby.gemspec